Pedro Leandro La Rotta
Machine Learning Engineer · Brooklyn, NY
MLE at JPMorganChase focused on Applied AI for payments. Previously worked on NER models at dMetrics. Background in robotics and control from MIT and Columbia University, where my MS research explored EMG-based learning systems for hand exoskeleton control. Originally from Bogotá, Colombia; usually found drinking coffee, listening to salsa, or out on a run.
Experience
-
Machine Learning Engineer · JPMorganChase
-
Machine Learning Engineer · dMetrics NLP systems.
-
Graduate Research Assistant · Columbia ROAM Lab EMG-based deep learning to control robotic hand exoskeletons.
-
Engineering Intern · Fermilab Automation for particle-accelerator target & SRF systems, as a GEM Fellow.
-
Research Associate · Octant Drug-discovery research.
-
Undergraduate Researcher · MIT ML for single-cell RNA-seq analysis (Lauffenburger Lab).
Research
- Meta-Learning for Fast Adaptation in Intent Inferral on a Robotic Hand Orthosis Few-shot MAML for surface-EMG intent classification that recalibrates from a handful of examples per session.
- ChatEMG: Synthetic Data Generation to Control a Robotic Hand Orthosis A generative model for EMG signals that bootstraps intent classifiers when real data is scarce.
Projects
- prism A sub-quadratic embedding model built on bidirectional recurrence.
- dreamformer Forcing explicit long- and short-term memory structure into a transformer.
- fca An explicit “future belief” branch added to a transformer to improve planning.
- agentbus A ROS-like harness for orchestrating multi-agent systems.
- kiki A personal memory unit for your agents.
- arxiv-digest A daily digest of the top papers in ML, NLP, and robotics.
- NumpyNet A PyTorch-free deep learning framework, built from scratch in NumPy.
- GeneTSP Evolving solutions to the traveling salesperson problem with genetic algorithms.